Aqua Licentiate Theses 2013:2
species? In this thesis, observed patterns of habitat use and population structures of three fish species common in lakes have been explored. Results from predation and competition experiments could not fully explain the habitat use of competing species. However, when also incorporating metabolism, the estimated net energy gain of competitors could increase the understanding of observed patterns in nature. Ulrika Beier received her Bachelor of Science degree, with a major in limnology, at Lund University. She has since 1990 worked within various fields concerning aquatic ecology in lakes and running waters, environmental monitoring and assessment, as well as fisheries related research at the Institute of Freshwater Research, Department of Aquatic Resources, SLU. Aqua Licentiate Theses presents licentiate theses from the Department of Aquatic Resources, Swedish University of Agricultural Sciences. SLU generates knowledge for the sustainable use of biological natural resources. Research, education, extension, as well as environmental monitoring and assessment are used to achieve this goal. Online publication of thesis summary: http://epsilon.slu.se/ ISBN 978-91-576-9155-2 (print version) ISBN 978-91-576-9156-9 (electronic version)
Aqua Licentiate Theses 2013:2 • Habitat Selection and Indirect Interactions in Fish Communities • Ulrika Beier
Which underlying mechanisms can explain the habitat distribution of
Habitat Selection and Indirect Interactions in Fish Communities Mechanisms to Explain Spatial Distribution of Perch, Roach, and Vendace Ulrika Beier
Licentiate Thesis Swedish University of Agricultural Sciences Department of Aquatic Resources Uppsala
Habitat Selection and Indirect Interactions in Fish Communities Mechanisms to Explain Spatial Distribution of Perch, Roach, and Vendace
Ulrika Beier Faculty of Natural Resources and Agricultural Sciences Department of Aquatic Resources Skolgatan 6, 742 42 Öregrund
Licentiate Thesis Swedish University of Agricultural Sciences Department of Aquatic Resources Uppsala 2013
Aqua Licentiate Theses 2013:2
Cover: After drawing by Görel Marklund (reproduced with permission)
ISBN 978-91-576-9155-2 (print version) ISBN 978-91-576-9156-9 (electronic version) © 2013 Ulrika Beier, Uppsala Print: SLU Service/Repro, Uppsala 2013
Habitat Selection and Indirect Interactions in Fish Communities. Mechanisms to Explain Spatial Distribution of Perch, Roach, and Vendace Abstract To increase the understanding of freshwater lake ecosystems, I have studied the habitat selection of perch (Perca fluviatilis L.), roach (Rutilus rutilus (L.)), and vendace (Coregonus albula (L.)). These fish species use the pelagic and the littoral-benthic habitats in lakes to different extents. Perch and roach are omnivorous, and perch become piscivorous at larger sizes. Vendace is a pelagic species specialized in eating zooplankton. Vendace was expected to affect biotic interactions and habitat use of roach and perch, both directly and indirectly. I used monitoring data to examine how species distribution patterns, as well as population structures, depended on species composition. In a predation experiment, I studied the relative predation sensitivity as well as evasive behaviours of roach and vendace, with piscivorous perch used as predators. In foraging experiments in aquaria, I studied foraging efficiencies and swimming performances of roach and vendace eating zooplankton in different temperature and light treatments. I then applied metabolic models for roach and vendace, respectively, to compare their net energy gain in different abiotic conditions. Roach used the pelagic habitat less, and the biomass of roach was lower in lakes with vendace. Results did not support the prediction that perch populations would benefit from the presence of vendace. However, results indicated that a release of competition for small perch may be mediated by vendace, through changed habitat use of roach, increasing the possibilities for predation. Roach and vendace were similar in their sensitivity to predation, indicating that energy gain can explain their habitat use. Foraging efficiencies did not explain the habitat use of roach and vendace in the field. However, the net energy gain in different abiotic conditions, could explain observed patterns of their habitat use in lakes. This thesis shows how the trade-off between mortality and net energy gain is manifested in habitat use. Including habitat selection in ecological studies may increase our understanding of biotic interactions. Metabolic costs as well as foraging abilities in different abiotic conditions are important for explaining the habitat use of species. Such knowledge can make it possible to forecast how interacting fish species may be affected by environmental change. Keywords: active metabolic rate, net energy gain, size-dependent interactions, environment, foraging efficiency, competition, predation Author’s address: Ulrika Beier, Department of Aquatic Resources, Institute of Freshwater Research, Stångholmsvägen 2, SE-178 93 Drottningholm, Sweden. E-mail:
[email protected]
Dedication To the memory of my grandmothers Saga and Inga
“Tho' much is taken, much abides; and though We are not now that strength which in old days Moved earth and heaven; that which we are, we are; One equal temper of heroic hearts, Made weak by time and fate, but strong in will To strive, to seek, to find, and not to yield.” Alfred Tennyson. Quote from “Ulysses”. Poems. London: Edward Moxon, Dover Street. 1842.
Contents List of Publications
7
Word list including abbreviations
9
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7
Introduction Habitat selection in ecological studies Biotic interactions influence habitat selection Abiotic factors define niches Metabolism is affected by abiotic factors The study system Species distribution patterns Underlying mechanisms for habitat selection
11 11 12 13 14 14 17 19
2
Objectives
21
3 3.1 3.2
Materials and methods Habitat use of perch, roach and vendace Underlying mechanisms for habitat distribution patterns
23 23 24
4 4.1 4.2 4.3 4.4
Results and discussion Habitat distribution patterns Effects on population structures in perch and roach The effect of predation Net energy gain explains habitat use of competitors
27 27 30 34 35
5
Conclusions and future perspectives
39
6
Summary
43
7
Sammanfattning
45
References
49
Acknowledgements
55
6
List of Publications This thesis is based on the work contained in the following papers, referred to by Roman numerals in the text: I
Beier, U. (2001). Habitat distribution and size structure in freshwater fish communities: effects of vendace on interactions between perch and roach. Journal of Fish Biology 59(6), 1437–1454.
II
Beier, U. (2013). Winning the loser’s game: implications of temperature and light for habitat selection of roach and vendace (submitted manuscript).
Paper I is reproduced with the permission of the publisher.
The contribution of U. Beier to the papers in this thesis was as follows: I
From an idea of U. Beier to study the distribution of common species among different habitats in lakes, the study was designed and planned jointly with L. Persson and M. Appelberg. U. Beier processed selection of data, conducted statistical analyses, and compiled the manuscript.
II
From the ideas of L. Persson, experiments were designed and planned jointly with L. Persson and M. Appelberg. U. Beier executed behavioural experiments, processed data, conducted statistical analyses, and compiled the manuscript. 7
8
Word list including abbreviations epilimnion ectothermic foraging habitat hypolimnion littoral-benthic metalimnion omnivore pelagic piscivorous predation thermocline
trophic level zooplankton
above the thermocline* in summer organisms which do not generate body heat; i.e., their bodies hold the same temperature as the surroundings searching for food and eating spatially defined life-space, characterized by physical or chemical factors, for example temperature or substrate below the thermocline* the parts of a lake close to shores and along the lake bottom thermocline* organism eating from different trophic levels* the parts of a lake consisting of open volumes of water without physical structure fish eating when one animal eats another distinct and limited depth interval in a column of fluid (e.g., water) in which temperature changes more rapidly with depth than it does in the layers above or below position in a food chain – primary producers are level one, herbivores are level two, etc. in the following text referring to miniature crustacean animals, filtering green algae or smaller animals
* word or abbreviation explained in the list
9
10
1
Introduction
1.1 Habitat selection in ecological studies Ecological communities are normally complex by nature. Community studies may be simplified by specifying different habitats within the spatial borders of the community. Habitats have been described as ‘infinite patches’, where the renewal of resources is in the same magnitude as the rate of gain of foragers (Stephens & Krebs, 1986). The theory of habitat selection concerns mechanisms for the organisms’ specialisation and choice of habitat, and resulting patterns in growth, survival and reproduction (Svärdson, 1949; Fretwell & Lucas 1970; Holt, 1984; Gilliam & Fraser, 1987, 1988; Rosenzweig, 1987; Bernstein et al., 1988; Brown, 1988; Morris, 1988; Pulliam & Danielson, 1991). One habitat may be separated from other habitats by different abiotic factors, such as light, temperature, or physical substrate and structure. Habitats can also be separated by biotic factors, such as types and amounts of prey and competitors, as well as predator abundance (Southwood, 1977). If there is a flow of one or several populations between habitats, the ecology of the community will depend not only on abiotic and biotic factors which define the habitats themselves, but also on relative amounts of the different habitats in the ecosystem (Morris, 1988; Oksanen, 1990; Pulliam & Danielson, 1991). In addition to being differently adapted for consuming different prey, organisms are also differently adapted to abiotic factors and their variability. These adaptions will often imply that different species are more or less inclined to choose certain habitats. A framework of morphological, physiological and behavioural adaptions is known as the ecological niche. The fundamental niche of a species is the wider spectrum of abiotic and biotic factors where the species can persist, whereas the realised niche is the actual, reduced spectrum in an ecosystem limited by geographical factors and intra- and inter-specific interactions. 11
1.2 Biotic interactions influence habitat selection The strength and magnitude of biotic interactions may change the distribution of populations among different habitats (Svärdson, 1949; Werner et al., 1983b; Gilliam & Fraser, 1987; Rosenzweig, 1987; Brown, 1988; Morris, 1988). A reason for this is density-dependent habitat selection, as individuals will move from a more crowded habitat to another less crowded, if it is more profitable per individual (Fretwell & Lucas 1970). Natural selection will favour individuals that maximize fitness when choosing between habitats. Individuals will therefore try to maximize energy gain and minimize predation risk (Cerri & Fraser, 1983; Gilliam & Fraser, 1987). The capacities to forage as well as to escape predators are size dependent, making the interactions between predators and prey depend on their relative sizes (Persson, 1988; Polis, 1988). Thus, in size-structured populations, such as characterize fish species, distribution among habitats is governed by sizedependent trade-offs between growth and mortality (Werner et al., 1983a; Werner & Gilliam, 1984; Fraser & Gilliam, 1987; Gilliam & Fraser, 1987). As fish continue to grow throughout their life, biotic interactions in fish communities will then depend on the ontogenetic niche-shifts that fish go through, i.e., how and when fish change food sources and/or habitat, from the origin throughout development, during their life span. These ontogenetic shifts will interact with the size structures of populations in the ecosystem. The cost of foraging can also be influenced by the physiological status of organisms (Taylor, 1984; Godin & Smith, 1988). This has been exemplified by Godin and Smith (1988), and Jakobsen et al. (1988), who found that risk-taking behaviour increased with hunger or energy deficit. Larger sizes of many fish species are omnivorous (predators on more than one trophic level), which implies that individuals can, depending on their relative sizes, be prey, competitors or predators to others (Polis, 1991). Furthermore, if organisms go through ontogenetic niche shifts by changing their diet or habitat choice during different size stages in their life cycle, a changed situation in one life stage may have consequences for the population itself, as well as for the structure and dynamics of other populations (Ebenman & Persson, 1988). Therefore, interactions within and among size-structured populations with ontogenetic niche shifts will affect the whole fish community. Biotic interactions affecting community structure can be both direct and indirect (Kerfoot & Sih, 1987; Strauss 1991). Direct interactions in a food web are the consumption of prey by a predator. Indirect interactions on the other 12
hand are the effects one species have on the interactions between a second and a third species (Miller & Kerfoot, 1987; Strauss 1991). Competition based on two species exploiting a resource can be seen as an indirect interaction with negative effects for both species (Abrams, 1987; Levitan, 1987). Furthermore, a predator which reduces the population of one prey species may have indirect positive effects on a) the species competing with the prey species (Abrams, 1987), or b) the species which themselves are subject to predation from the prey species (Vanni, 1987). A chain of indirect interactions is also known as a trophic cascade (Carpenter et al,. 1985). A negative indirect effect of predation is “apparent competition” (Holt, 1977), meaning that when two species share a common predator, an increase in one of the prey species causes a decrease in the other, as the predator population may benefit from the increase in the first species and thereby indirectly exerting more predation pressure also on the other species. A complicating factor in an effect chain may be if the direct or indirect interactions are asymmetric, i.e. when two competing species are different in how efficient they are at exploiting a resource (Persson, 1988). Because energy gain and mortality determine the net profitability of a habitat, altered competition and predation situations may affect habitat selection and thereby change the distribution of species, or size groups within species, between habitats (Werner et al,. 1983a; Gilliam & Fraser, 1987; Brown, 1988). Thus, direct and indirect interactions may be mediated by altered habitat use, and may also affect the habitat use of species.
1.3 Abiotic factors define niches Abiotic factors are the base of existence for species, forming the structure and function of ecological communities (Dunson & Travis, 1991). Temperature, light and oxygen levels at different depths are important factors for fish. Temporal heterogeneity of the environment is another factor affecting communities (Menge & Sutherland, 1976). Temperature is a geographical and physical factor to which fish species are differently adapted. Being ectothermic organisms, fish are depending on the surrounding temperatures for metabolic activity, which allows for mobility and somatic growth. However, the total energy cost increases with temperature, as an increased activity level in higher temperatures also leads to higher energy expenditure. Besides temperature, visually hunting fish are highly depending on their sense of vision to find food (Guthrie & Muntz, 1993). Species may be differently adapted to different light intensities, which may affect their relative 13
competitive abilities (Bergman, 1988; Diehl, 1988). This is particularly important for fish communities in lakes, where the light regime depends on time of season, time of day, as well as the water depth. Additionally, water colour and turbidity affects the light climate in the water column (Ranåker et al., 2012), which may differently affect behaviours of different fish species (Guthrie & Muntz, 1993; Ranåker, 2012).
1.4 Metabolism is affected by abiotic factors Metabolism is a principal force in ecology, linking e.g. temperature to the ecology of populations and whole communities (Brown et al., 2004). As an example of different metabolic adaptions to different temperatures, salmonid fish have a higher active metabolic rate than cyprinid fish at 15°C, which is approximately the temperature optimum of many salmonid species (Johnston & Dunn, 1987; Clarke & Johnston, 1999). Several studies have shown that cyprinid fish can instead benefit from temperatures warmer than 15°C (Persson et al., 1991; Holmgren & Appelberg, 2000; Graham & Harrod, 2009). It has been suggested that energy costs caused by activity, e.g. swimming, might be one of the most important factors for understanding amongpopulation variability in fish growth rates (Boisclair & Leggett, 1989). In connection to adaptions to different temperatures, average swimming speeds at different temperatures can differ depending on the morphology, metabolism and functional specialization of the species. As swimming activity infers an energetic cost, it is important to recognize in connection with foraging efficiency, to be able to understand the mechanisms for habitat selection and the competitive abilities of different species. Metabolic traits as well as costs of moving can together with foraging efficiency also help to understand the distribution of species at a larger scale. Metabolic costs often also depend on body size (Clarke & Johnston, 1999). Moreover, how metabolic costs change with temperature and body size differ between species (Ohlberger et al., 2012). As a consequence, mechanisms for habitat selection may vary between species, and between differently sized individuals.
1.5 The study system There is an advantage in using lakes in ecological studies, as each lake constitutes a sample of a semi-closed ecosystem, thus enabling the use of many samples when studying ecosystem functions (Schindler, 1990). In Scandinavia there is a further advantage in the availability to a great number of lakes, 14
however individually unique. In lakes, two distinct habitats may be defined: the littoral-benthic zone and the pelagic zone. The littoral-benthic zone extends from shallow, near-shore areas, next to the bottom all over the lake. This is a heterogeneous habitat in terms of physical structure, as well as numbers and sizes of available prey types. The pelagic zone is constituted by the open water volume away from the shore, where the water is deeper. This is more homogenous, although available food is often aggregated vertically and horizontally. Studying the distribution of roach (Rutilus rutilus (L.)), perch (Perca fluviatilis L.), and vendace (Coregonus albula (L.)) in these two habitats can then illustrate how habitat selection may be governed by sizedependent trade-offs between growth and mortality (Fig. 1).
PP
PC ReL
RC
VC ReP
Figure 1. Simplified food-web of the three focal fish species and their resources in two habitats. PP is piscivorous perch, PC, RC, and VC are perch, roach and vendace competing in one or both habitats. ReL and ReP are the food resources in the littoral-benthic and pelagic, respectively. —, Littoral-benthichabitat; – – –, pelagic habitat.
15
Roach and perch are widespread, and often the two numerically dominant fish species in Scandinavian lake systems. Roach is an efficient zooplanktivore potentially shifting to larger invertebrates, but is also able to use algae and detritus as a food source (Hellawell, 1972; Persson, 1983a). Roach are mainly found in littoral-benthic areas, but earlier studies have documented that roach perform horizontal migrations out to the pelagic zone at night (Bohl, 1980; Gliwicz & Jachner 1992). Perch has a life-history including shifts in diet and habitat (Persson, 1983b). As it grows, perch will shift from feeding on zooplankton to macroinvertebrate prey, and eventually to fishes (Alm, 1946; Craig, 1974; Nyberg, 1979). Shortly after hatching in the littoral zone, perch fry move to the pelagic zone to feed on zooplankton (Byström et al., 2003), and at a size of 10-30 mm shift back to mainly using the littoral-benthic zone (Treasurer, 1988; Wang & Eckmann, 1994; Byström et al., 2003). The habitat use of intermediate-sized and large perch is density-dependent and variable, but in general, perch are mainly found in or close to the littoral-benthic zone of lakes where a large variety of food items, including small fish, is available for this omnivorous species (Horppila et al., 2000; Kahl & Radke, 2006). The competitive relationship between small perch and roach can eventually be replaced by predation from perch, as perch will change food source and eventually begin to eat fish at a certain size; i.e., perch is an example of a an ontogenetic omnivore (Persson et al., 2000). The asymmetric competitive relationship between roach and perch is a well-studied example of where the ontogenetic omnivory in perch is an important mechanism behind dominance relationships between these two species (Persson, 1983c; Persson, 1988). As roach are more effective predators on zooplankton than perch, the degree of interspecific competition between the two species for this resource will limit the proportion of the perch population reaching the piscivorous stage (Persson, 1986, 1987; Persson & De Roos, 2012; Persson & Greenberg, 1990; Persson et al., 1992). On the other hand, if piscivorous perch are present and are able to reduce populations of smaller planktivorous fishes, the competitive pressure experienced by non-piscivorous perch may be reduced (Persson 1983c, 1986b; Persson and De Roos 2012; Johansson & Persson, 1986; Svanbäck & Persson, 2004). If perch individuals will begin to eat fish at intermediate sizes, they will normally grow faster (Claessen et al., 2000; Le Cren, 1992; Persson et al., 2000). Piscivorous perch with fast individual growth will consume more fish prey. This situation may initiate a causal loop, where competing prey fishes are consumed to an extent that individual smaller perch will grow faster as a result of reduced competition, and they can in turn more easily switch to piscivory, reinforcing the feedback between individual growth and biotic interactions between perch and roach. 16
To explore how altered biotic interactions may affect the distribution of roach and perch populations among different habitats, the study system will include vendace, which has a strong preference for the pelagic habitat. Vendace is highly specialized for preying on zooplankton during its entire life cycle (Hamrin, 1983; 1986). Based on morphology, i.e., a protruding lower jaw and a high number of gill rakers, vendace is expected to be the superior competitor of the three species in the pelagic habitat (Svärdson, 1976; Hamrin, 1986). Vendace may also exploit a temperature refuge in the pelagic zone and forage at low temperatures and at low light levels in the hypolimnion (Northcote & Rundberg, 1970; Dembinski, 1971; Hamrin & Persson, 1986; Persson et al., 1991; Mehner et al., 2007). There are indications that vendace may counteract the effect of increased lake productivity, which normally favours roach before perch (Persson et al., 1991; 1992). Piscivorous perch might be favoured in systems with vendace, which would have consequences for all their fish prey populations (Appelberg et al., 1990; Persson et al., 1991; 1992). This can be explained by effects which vendace might have on roach, by reducing the common food resource consisting of zooplankton in the pelagic habitat, and/or through apparent competition. Apparent competition may here be caused by vendace constituting an alternative prey for piscivorous perch, and thereby indirectly increasing predation pressure for roach, as a larger proportion of the perch population may become fish eating. Alternatively, or additionally, it may be an indirect effect of altered habitat use; increased competition may affect the habitat use of roach, forcing roach to increase its use of the littoral zone where there may be more piscivorous perch, leading to that roach may be subdued to a larger predation risk. This situation would describe apparent competition mediated by altered habitat use, which is one example of where indirect interactions are linked with habitat selection.
1.6 Species distribution patterns The conclusions in earlier studies on the positive effects of vendace on perch, and thereby on the habitat distribution of perch, vendace and roach, were based on a relatively low number of lakes, which were situated in a limited geographical area (Appelberg et al., 1991; Persson et al., 1991, Figure 2). In Paper I, I therefore explore data from a larger number of lakes (N=115) within a widespread geographical area in Sweden (Figure 2). The aim of Paper I was to allow for detailed testing of hypotheses on mechanisms explaining patterns found in earlier studies of species distributions and population structures (Appelberg et al., 1991; Persson et al., 1991), concerning the habitat use and 17
possible indirect effects of vendace on perch and roach (see Materials and methods). I used fish sampling data from both littoral and pelagic habitats in lakes to test the generality of earlier findings. Data were collected from the NORS database containing standardized test fishing data from national and regional monitoring programmes (Kinnerbäck, 2013). Fishing occasions where multi-mesh gillnets had been used in both the littoral-benthic and pelagic habitats were selected.
!
!
D D !! ! ( ! ! (
D
! (
! D ! D
DD !! D !! ! ( D DD D D ! D! !D D ( ! D !! !( ( D D D !D ( ! D( ! ( !! ! D( ! ( D ! ! !! !( !! ! D ! ( D! !( ( ! !D !! !( ( ! !( ( ! D ! ! !D D ! !!(( ! D ! !!( ! ( ! ! D D! ( ( ! !!!( ( ! ( ! !D !!
( !
( !! ! ! ! ( ! ! ! ( ! ! ( ( !
Figure 2. Map of Sweden with locations of lakes used in Paper I (lakes with perch = , with perch and roach = , and with perch, roach and vendace = ○), and in the study of (Persson et al., 1991) (red circles).
18
1.7 Underlying mechanisms for habitat selection Using experiments both to investigate sensitivity to predation, and relative competitive abilities as well as energetic gain in competing species, will aid in understanding the relative importance of these factors when looking for explanations for the habitat distribution patterns in the field. In Paper II, I therefore performed both predation and feeding experiments on roach and vendace to study the mechanisms underlying their trade-off in mortality risk to energy gain in different habitats. There are several examples of earlier studies of biotic interactions and behaviour in fish using enclosure experiments with semi-natural conditions (Werner et al., 1983a, b; Eklöv & Persson, 1995; 1996). By performing experiments in pond enclosures, it is normally possible to both study natural behaviour of the organisms in focus, as well as quantitatively measure consumption of different prey items. I therefore used perch predators in an open water environment to study the relative sensitivities of roach and vendace to predation, as well as their evasive behaviour (Paper II). Individual foraging abilities and the effects of environmental factors may efficiently be studied in controlled environments using laboratory experiments (Persson, 1988; Bergman, 1988; Diehl, 1988). To better understand mechanisms underlying competition and patterns of relative abundance of roach and vendace across different lake habitats, I therefore designed laboratory experiments where temperature and light were used as abiotic factors (Paper II). The relative efficiency of vendace as a zooplankton predator compared to e.g. roach has so far not been quantified, although metabolic information has been collected for both species separately (Hölker, 2003; 2006; Ohlberger et al., 2007). I used experiments in aquaria to investigate the relative foraging capacities in different temperature and light treatments, as well as collecting swimming speed data, to calculate estimates of species- and sizespecific metabolism as well as net energy gain in different temperatures (Paper II).
19
20
2
Objectives
Increased understanding of ecological interactions and habitat preferences of species can be used to predict how changes in the environment may affect species specifically, which in turn is necessary to be able to foresee the effects of management measures or other human impact on fish communities. The goal with this thesis is to increase the knowledge and understanding of freshwater lake systems by studying the habitat selection of three common fish species in lakes: perch, roach, and vendace. I examine patterns of habitat distribution and population dynamics observed in lakes, and experimentally study biotic interactions between these three species, searching to explain the mechanisms by which the outcome of these interactions are manifested through habitat selection. The major questions I address in this thesis are:
How do habitat distribution patterns differ between the species? Do the population structures of perch and roach differ depending on whether vendace is present or not? If roach and vendace are the main species occurring in the pelagic habitat, what are the mechanisms that can explain their habitat use? How can the trade-off between risks of being eaten versus energy intake be understood by these mechanisms? Can the mechanisms found in the study further help in understanding the habitat distribution of species?
21
22
3
Materials and methods
3.1 Habitat use of perch, roach and vendace To investigate the habitat distribution of the three studied fish species in two habitats, I used data from 115 lakes (Paper I). Size structures as well as relative abundance in the littoral-benthic and pelagic habitats were compared among lakes having: only perch; perch and roach; or perch, roach and vendace. Based on results from previous studies regarding competitive abilities of roach and perch as well as population dynamics (Svanbäck & Persson, 2004), four hypotheses were tested: (1) Habitat choice in roach will be a trade-off between foraging gain and risk of predation. Thus, the pelagic habitat will be less used by zooplanktivorous perch and roach in lakes with vendace, as a consequence of resource competition. As a consequence of trade-offs between growth and mortality rates, habitat choice will also be size dependent so that the sizes of roach found in the pelagic zone will be within the range where foraging on zooplankton is profitable, and where the risk of predation in an open habitat is significantly reduced. (2) The relative biomass of perch in the pelagic habitat will be related to the biomass of zooplanktivores (a positive relation for the biomass of piscivorous perch and a negative relation for non-piscivorous perch). (3) The proportion of piscivorous perch will be higher in perch-roachvendace lakes than in perch-roach lakes, because of a higher growth for perch in lakes with vendace as an additional prey. (4) Roach size distributions will be skewed towards larger sizes in lakes with vendace compared to lakes without vendace, as piscivorous perch are expected to be more abundant in perch-roach-vendace lakes (hypothesis 3). This is expected to lead to a higher mortality for small-sized roach from perch predation, resulting in a higher proportion of larger roach, as larger size confers a refuge from predation by perch. 23
3.2 Und derlying mechanisms fo or habitat diistribution p atterns To study the t relative sensitivity of ro oach and vend dace to predatiion from perch in an open n water envirronment, I peerformed pred dation experim ments in ponnd enclosuress lacking vegeetation, with mean m water deepth 1.1 m (F Figure 3, Papeer II). I used perch as pred dators and eith her roach or vendace v as preey, as well as a w both speccies. Behaviou ur of predatorrs and evasivve mixed preey treatment with behaviour of prey, as well w as swimm ming speeds fo or the differennt species werre A two day ys, the remain ning prey fish were collecteed and counted recorded. After to determiine the captu ure success off perch in th he different trreatments witth different fiish prey. As field d data show th hat vendace raarely use the littoral-benthi l ic zone and arre more comm mon in deeperr, darker wateer in the pelag gic zone (Ham mrin & Perssonn, 1986; Perssson et al., 19 991; Mehner et al., 2007, Paper P I), wheere they would have a preedation refugee from perch, I predicted th hat vendace w would be morre susceptiblee than roach to o predation fro om perch.
Figure 3. Ph hotograph showin ng pond enclosu ures with the obsservation tower, used in predatioon experiments (Paper II).
24
To study competition between roach and vendace, I performed laboratory experiments in aquaria, measuring relative foraging capacities as well as swimming speeds of roach and vendace in different temperature and light conditions (Figure 4, Paper II). In the experiments, roach and vendace were foraging on different densities of zooplankton to investigate their functional response under varying light and temperature conditions. Furthermore, I used the recorded swimming speed data in bioenergetic models for each species, respectively, to test whether estimated net energy gain could explain the habitat distribution of the two species in the field (Paper II).
A
B
Figure 4. Schematic picture of aquarium experiment (Paper II). A) Before fish were released, and B) recording the capture rate and swimming speed of one randomly selected individual when eating zooplankton (Daphnia magna) during the experiments.
In the laboratory experiment, I used a combination of natural temperature and light conditions in thermally stratified lakes, to be able to elucidate possible mechanisms explaining competition-driven patterns of habitat use as well as relative abundance of roach and vendace. The size of fish in the foraging experiments matched the predominant size interval of roach found in the pelagic zone in lakes. I chose temperature treatments according to a standard situation in temperature stratified lakes during summer (Paper II). At a temperature corresponding to the epilimnion (above the thermocline), where the two species coexist in the pelagic zone of lakes, I used two different light treatments. I chose light treatments as to resemble normal light levels in the epilimnion during daylight, and during dusk or dawn. 25
I predicted that both species increase capture rates as well as swimming speeds with increased temperature. Based on spatial distribution patterns observed in lakes, where vendace is normally found in deeper (Northcote & Rundberg, 1970; Dembinski, 1971; Hamrin & Persson, 1986; Persson et al., 1991; Mehner et al. 2007; Paper I), and thus colder and darker water than roach, I predicted that a) the capture rate of vendace would be less affected by low light levels than it would be for roach, b) vendace would have higher metabolic costs compared to roach in warmer waters, and c) the net energy gain of vendace would be higher than for roach at lower temperatures, while d) the net energy gain for roach at the highest temperature would be higher, than for vendace.
26
4
Results and discussion
4.1 Habitat distribution patterns The distribution of the three species differed between the pelagic and littoralbenthic habitats, and depended on the presence of the other species (Paper I). Roach biomass was lower in lakes with vendace, and roach used the pelagic habitat to a relatively less extent in the presence of vendace (Figure 5a). The relative biomass in the pelagic habitat, i.e. the ratio of pelagic biomass to littoral-benthic biomass, was significantly lower for roach in the presence of vendace (Figure 5b). This suggests that competition from the specialized zooplanktivorous species vendace has a clear negative effect on roach. Furthermore, roach mainly used the 0-6 m depth interval in both the pelagic zone and the littoral-benthic zone, and were least common in the pelagic zone below 6 m. Vendace were found mostly below 6 m in the pelagic zone, which indicates that vendace, in contrast to roach are able to explore the zooplankton resource in the deeper part of the pelagic zone (Paper I). As zooplankton can perform diel horizontal migrations and move into deeper waters to avoid predation (Lampert, 1993), the predation pressure for zooplankton may be stronger when vendace is present, as a refuge for zooplankton in deeper water might be lacking. In support of this view, vendace has been shown to strongly deplete the zooplankton resource (Helminen & Sarvala, 1997), which suggests a potential strong effect of their zooplankton consumption for competing species in the pelagic zone. In this study, relative biomass of perch in the pelagic habitat was higher in lakes with zooplanktivores, i.e., only roach, or both roach and vendace, compared to lakes with only perch present (Paper I, Figure 5b). This may be explained by intra-specific competition for perch in the littoral-benthic habitat in lakes with roach, or in lakes with both roach and vendace, leading to that perch use the pelagic zone to a greater extent with competing species present. 27
This possible explanation is supported by results from Svanbäck et al., (2008), who suggested that intra-specific competition was more important than interspecific competition for the habitat use of perch. It can be expected that interspecific competition from other zooplanktivorous species should also increase intra-specific competition for non-piscivorous perch in the littoral-benthic zone by reducing the available zooplankton resource. Increased competition in the littoral-benthic habitat may explain why perch use the pelagic habitat to a greater extent when competing species are present in the system. When comparing the lake group with perch and roach to the lake group with perch, roach and vendace, there was no significant difference in the ratio of pelagic perch (total) biomass to littoral-benthic perch (total) biomass (ANOVA, P=0.390). Furthermore, there were no significant differences in the ratio of pelagic to littoral-benthic biomass of piscivorous (ANOVA P=0.568), or non-pisccivorous perch (ANOVA, P=0.290) between lake groups. For nonpiscivorous perch, this pattern may primarily be explained by relative foraging efficiencies on the zooplankton resource, where roach is more efficient than perch (Persson 1983c). Competition from roach may then lead to that perch would gain less energy in the pelagic zone. In this case, the pelagic habitat would already be less profitable for perch even if only roach were present. Therefore, it is unlikely that the additional effect of vendace as a superior competitor in the pelagic zone would be clearly manifested regarding the habitat use of non-piscivorous perch (Paper I). This may be a result of an indirect effect of increased competition in the littoral-benthic habitat. If roach are forced out from the pelagic zone by the presence of vendace, thus increasing the competition in the littoral-benthic zone, this may lead to that the relative profitability of the two habitats may be similar for non-piscivorous perch, irrespective of whether vendace is present or not. For piscivorous perch, the lack of differences between lake groups in habitat use may indicate indirect interactions resulting in altered habitat use of suitable prey fish. Because of competition from vendace, resulting in altered habitat use of roach and/or non-piscivorous perch, which may in turn affect availabilities of prey fish, could lead to a similar relative profitability for piscivorous perch in both habitats. However, to understand the relative habitat profitabilities for perch depending on the presence of vendace would require detailed studies of food availability and perch diets in different systems.
28
Figure 5. (a) Biomasses ± 1 S.E. in the littoral-benthic and pelagic habitats of three groups of lakes. P, perch; PR, perch and roach; PRV, perch, roach and vendace. Top bar in each lake group=perch biomass; middle bar=roach and lowest bar=vendace. For perch, darker shading indicates the piscivorous proportion of the biomass. (b) Mean ratios ± 1 S.E. of the total biomass of non-piscivorous perch, piscivorous perch and roach in the pelagic to that in the littoral-benthic habitat. Black bars represent lakes with perch, grey staples lakes with perch and roach, and white staples represent lakes with perch, roach and vendace.
29
4.2 Effects on population structures in perch and roach Results from the field study (Paper I) shows that the biomass of piscivorous perch in both the littoral-benthic and the pelagic habitats was positively related to the biomass of zooplanktivorous fish, but when controlling for total biomass, the relationship was not significant. Thus, in contrast to predictions, neither the relative amount of perch piscivorous biomass, nor the non-piscivorous perch biomass showed any significant relationship to the biomass of roach and vendace, i.e., when controlling for total biomass. Persson et al. (1991) predicted that the higher proportion of piscivorous perch, which they found in lakes with vendace, could be explained by indirect biotic interactions favouring perch populations. The arguments regarding relative habitat qualities for perch, with suggested indirect effects of habitat utilization by foremost roach (see Section 4.1), as well as the assumption that intra-specific interactions are important for perch (Svanbäck et al., 2008), are supported by observed patterns of individual growth in perch (Paper I). In the smallest size class (
